[PATCH] VFS: Make filldir_t and struct kstat deal in 64-bit inode num…

…bers

These patches make the kernel pass 64-bit inode numbers internally when
communicating to userspace, even on a 32-bit system.  They are required
because some filesystems have intrinsic 64-bit inode numbers: NFS3+ and XFS
for example.  The 64-bit inode numbers are then propagated to userspace
automatically where the arch supports it.

Problems have been seen with userspace (eg: ld.so) using the 64-bit inode
number returned by stat64() or getdents64() to differentiate files, and
failing because the 64-bit inode number space was compressed to 32-bits, and
so overlaps occur.

This patch:

Make filldir_t take a 64-bit inode number and struct kstat carry a 64-bit
inode number so that 64-bit inode numbers can be passed back to userspace.

The stat functions then returns the full 64-bit inode number where
available and where possible.  If it is not possible to represent the inode
number supplied by the filesystem in the field provided by userspace, then
error EOVERFLOW will be issued.

Similarly, the getdents/readdir functions now pass the full 64-bit inode
number to userspace where possible, returning EOVERFLOW instead when a
directory entry is encountered that can't be properly represented.

Note that this means that some inodes will not be stat'able on a 32-bit
system with old libraries where they were before - but it does mean that
there will be no ambiguity over what a 32-bit inode number refers to.

Note similarly that directory scans may be cut short with an error on a
32-bit system with old libraries where the scan would work before for the
same reasons.

It is judged unlikely that this situation will occur because modern glibc
uses 64-bit capable versions of stat and getdents class functions
exclusively, and that older systems are unlikely to encounter
unrepresentable inode numbers anyway.

[akpm: alpha build fix]
Signed-off-by: David Howells <[email protected]>
Cc: Trond Myklebust <[email protected]>
Cc: Al Viro <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>

arch/alpha/kernel/osf_sys.c

@@ -111,22 +111,26 @@ struct osf_dirent_callback {
 
 static int
 osf_filldir(void *__buf, const char *name, int namlen, loff_t offset,
-	    ino_t ino, unsigned int d_type)
+	    u64 ino, unsigned int d_type)
 {
 	struct osf_dirent __user *dirent;
 	struct osf_dirent_callback *buf = (struct osf_dirent_callback *) __buf;
 	unsigned int reclen = ROUND_UP(NAME_OFFSET + namlen + 1);
+	unsigned int d_ino;
 
 	buf->error = -EINVAL;	/* only used if we fail */
 	if (reclen > buf->count)
 		return -EINVAL;
+	d_ino = ino;
+	if (sizeof(d_ino) < sizeof(ino) && d_ino != ino)
+		return -EOVERFLOW;
 	if (buf->basep) {
 		if (put_user(offset, buf->basep))
 			return -EFAULT;
 		buf->basep = NULL;
 	}
 	dirent = buf->dirent;
-	put_user(ino, &dirent->d_ino);
+	put_user(d_ino, &dirent->d_ino);
 	put_user(namlen, &dirent->d_namlen);
 	put_user(reclen, &dirent->d_reclen);
 	if (copy_to_user(dirent->d_name, name, namlen) ||

arch/ia64/ia32/sys_ia32.c

@@ -125,18 +125,23 @@ sys32_execve (char __user *name, compat_uptr_t __user *argv, compat_uptr_t __use
 
 int cp_compat_stat(struct kstat *stat, struct compat_stat __user *ubuf)
 {
+	compat_ino_t ino;
 	int err;
 
 	if ((u64) stat->size > MAX_NON_LFS ||
 	    !old_valid_dev(stat->dev) ||
 	    !old_valid_dev(stat->rdev))
 		return -EOVERFLOW;
 
+	ino = stat->ino;
+	if (sizeof(ino) < sizeof(stat->ino) && ino != stat->ino)
+		return -EOVERFLOW;
+
 	if (clear_user(ubuf, sizeof(*ubuf)))
 		return -EFAULT;
 
 	err  = __put_user(old_encode_dev(stat->dev), &ubuf->st_dev);
-	err |= __put_user(stat->ino, &ubuf->st_ino);
+	err |= __put_user(ino, &ubuf->st_ino);
 	err |= __put_user(stat->mode, &ubuf->st_mode);
 	err |= __put_user(stat->nlink, &ubuf->st_nlink);
 	err |= __put_user(high2lowuid(stat->uid), &ubuf->st_uid);
@@ -1222,24 +1227,28 @@ struct readdir32_callback {
 };
 
 static int
-filldir32 (void *__buf, const char *name, int namlen, loff_t offset, ino_t ino,
+filldir32 (void *__buf, const char *name, int namlen, loff_t offset, u64 ino,
 	   unsigned int d_type)
 {
 	struct compat_dirent __user * dirent;
 	struct getdents32_callback * buf = (struct getdents32_callback *) __buf;
 	int reclen = ROUND_UP(offsetof(struct compat_dirent, d_name) + namlen + 1, 4);
+	u32 d_ino;
 
 	buf->error = -EINVAL;	/* only used if we fail.. */
 	if (reclen > buf->count)
 		return -EINVAL;
+	d_ino = ino;
+	if (sizeof(d_ino) < sizeof(ino) && d_ino != ino)
+		return -EOVERFLOW;
 	buf->error = -EFAULT;	/* only used if we fail.. */
 	dirent = buf->previous;
 	if (dirent)
 		if (put_user(offset, &dirent->d_off))
 			return -EFAULT;
 	dirent = buf->current_dir;
 	buf->previous = dirent;
-	if (put_user(ino, &dirent->d_ino)
+	if (put_user(d_ino, &dirent->d_ino)
 	    || put_user(reclen, &dirent->d_reclen)
 	    || copy_to_user(dirent->d_name, name, namlen)
 	    || put_user(0, dirent->d_name + namlen))
@@ -1287,17 +1296,21 @@ sys32_getdents (unsigned int fd, struct compat_dirent __user *dirent, unsigned i
 }
 
 static int
-fillonedir32 (void * __buf, const char * name, int namlen, loff_t offset, ino_t ino,
+fillonedir32 (void * __buf, const char * name, int namlen, loff_t offset, u64 ino,
 	      unsigned int d_type)
 {
 	struct readdir32_callback * buf = (struct readdir32_callback *) __buf;
 	struct old_linux32_dirent __user * dirent;
+	u32 d_ino;
 
 	if (buf->count)
 		return -EINVAL;
+	d_ino = ino;
+	if (sizeof(d_ino) < sizeof(ino) && d_ino != ino)
+		return -EOVERFLOW;
 	buf->count++;
 	dirent = buf->dirent;
-	if (put_user(ino, &dirent->d_ino)
+	if (put_user(d_ino, &dirent->d_ino)
 	    || put_user(offset, &dirent->d_offset)
 	    || put_user(namlen, &dirent->d_namlen)
 	    || copy_to_user(dirent->d_name, name, namlen)

arch/mips/kernel/linux32.c

@@ -77,6 +77,8 @@ int cp_compat_stat(struct kstat *stat, struct compat_stat __user *statbuf)
 	memset(&tmp, 0, sizeof(tmp));
 	tmp.st_dev = new_encode_dev(stat->dev);
 	tmp.st_ino = stat->ino;
+	if (sizeof(tmp.st_ino) < sizeof(stat->ino) && tmp.st_ino != stat->ino)
+		return -EOVERFLOW;
 	tmp.st_mode = stat->mode;
 	tmp.st_nlink = stat->nlink;
 	SET_UID(tmp.st_uid, stat->uid);

arch/mips/kernel/sysirix.c

@@ -1739,12 +1739,13 @@ struct irix_dirent32_callback {
 #define ROUND_UP32(x) (((x)+sizeof(u32)-1) & ~(sizeof(u32)-1))
 
 static int irix_filldir32(void *__buf, const char *name,
-	int namlen, loff_t offset, ino_t ino, unsigned int d_type)
+	int namlen, loff_t offset, u64 ino, unsigned int d_type)
 {
 	struct irix_dirent32 __user *dirent;
 	struct irix_dirent32_callback *buf = __buf;
 	unsigned short reclen = ROUND_UP32(NAME_OFFSET32(dirent) + namlen + 1);
 	int err = 0;
+	u32 d_ino;
 
 #ifdef DEBUG_GETDENTS
 	printk("\nirix_filldir32[reclen<%d>namlen<%d>count<%d>]",
@@ -1753,12 +1754,15 @@ static int irix_filldir32(void *__buf, const char *name,
 	buf->error = -EINVAL;	/* only used if we fail.. */
 	if (reclen > buf->count)
 		return -EINVAL;
+	d_ino = ino;
+	if (sizeof(d_ino) < sizeof(ino) && d_ino != ino)
+		return -EOVERFLOW;
 	dirent = buf->previous;
 	if (dirent)
 		err = __put_user(offset, &dirent->d_off);
 	dirent = buf->current_dir;
 	err |= __put_user(dirent, &buf->previous);
-	err |= __put_user(ino, &dirent->d_ino);
+	err |= __put_user(d_ino, &dirent->d_ino);
 	err |= __put_user(reclen, &dirent->d_reclen);
 	err |= copy_to_user((char __user *)dirent->d_name, name, namlen) ? -EFAULT : 0;
 	err |= __put_user(0, &dirent->d_name[namlen]);
@@ -1837,7 +1841,7 @@ struct irix_dirent64_callback {
 #define ROUND_UP64(x) (((x)+sizeof(u64)-1) & ~(sizeof(u64)-1))
 
 static int irix_filldir64(void *__buf, const char *name,
-	int namlen, loff_t offset, ino_t ino, unsigned int d_type)
+	int namlen, loff_t offset, u64 ino, unsigned int d_type)
 {
 	struct irix_dirent64 __user *dirent;
 	struct irix_dirent64_callback * buf = __buf;

arch/parisc/hpux/fs.c

@@ -77,17 +77,21 @@ static int filldir(void * __buf, const char * name, int namlen, loff_t offset,
 {
 	struct hpux_dirent * dirent;
 	struct getdents_callback * buf = (struct getdents_callback *) __buf;
+	ino_t d_ino;
 	int reclen = ROUND_UP(NAME_OFFSET(dirent) + namlen + 1);
 
 	buf->error = -EINVAL;	/* only used if we fail.. */
 	if (reclen > buf->count)
 		return -EINVAL;
+	d_ino = ino;
+	if (sizeof(d_ino) < sizeof(ino) && d_ino != ino)
+		return -EOVERFLOW;
 	dirent = buf->previous;
 	if (dirent)
 		put_user(offset, &dirent->d_off);
 	dirent = buf->current_dir;
 	buf->previous = dirent;
-	put_user(ino, &dirent->d_ino);
+	put_user(d_ino, &dirent->d_ino);
 	put_user(reclen, &dirent->d_reclen);
 	put_user(namlen, &dirent->d_namlen);
 	copy_to_user(dirent->d_name, name, namlen);

arch/parisc/kernel/sys_parisc32.c

@@ -237,14 +237,19 @@ int sys32_settimeofday(struct compat_timeval __user *tv, struct timezone __user
 
 int cp_compat_stat(struct kstat *stat, struct compat_stat __user *statbuf)
 {
+	compat_ino_t ino;
 	int err;
 
 	if (stat->size > MAX_NON_LFS || !new_valid_dev(stat->dev) ||
 	    !new_valid_dev(stat->rdev))
 		return -EOVERFLOW;
 
+	ino = stat->ino;
+	if (sizeof(ino) < sizeof(stat->ino) && ino != stat->ino)
+		return -EOVERFLOW;
+
 	err  = put_user(new_encode_dev(stat->dev), &statbuf->st_dev);
-	err |= put_user(stat->ino, &statbuf->st_ino);
+	err |= put_user(ino, &statbuf->st_ino);
 	err |= put_user(stat->mode, &statbuf->st_mode);
 	err |= put_user(stat->nlink, &statbuf->st_nlink);
 	err |= put_user(0, &statbuf->st_reserved1);
@@ -312,16 +317,20 @@ filldir32 (void *__buf, const char *name, int namlen, loff_t offset, ino_t ino,
 	struct linux32_dirent __user * dirent;
 	struct getdents32_callback * buf = (struct getdents32_callback *) __buf;
 	int reclen = ROUND_UP(NAME_OFFSET(dirent) + namlen + 1, 4);
+	u32 d_ino;
 
 	buf->error = -EINVAL;	/* only used if we fail.. */
 	if (reclen > buf->count)
 		return -EINVAL;
+	d_ino = ino;
+	if (sizeof(d_ino) < sizeof(ino) && d_ino != ino)
+		return -EOVERFLOW;
 	dirent = buf->previous;
 	if (dirent)
 		put_user(offset, &dirent->d_off);
 	dirent = buf->current_dir;
 	buf->previous = dirent;
-	put_user(ino, &dirent->d_ino);
+	put_user(d_ino, &dirent->d_ino);
 	put_user(reclen, &dirent->d_reclen);
 	copy_to_user(dirent->d_name, name, namlen);
 	put_user(0, dirent->d_name + namlen);
@@ -371,12 +380,16 @@ fillonedir32 (void * __buf, const char * name, int namlen, loff_t offset, ino_t
 {
 	struct readdir32_callback * buf = (struct readdir32_callback *) __buf;
 	struct old_linux32_dirent __user * dirent;
+	u32 d_ino;
 
 	if (buf->count)
 		return -EINVAL;
+	d_ino = ino;
+	if (sizeof(d_ino) < sizeof(ino) && d_ino != ino)
+		return -EOVERFLOW;
 	buf->count++;
 	dirent = buf->dirent;
-	put_user(ino, &dirent->d_ino);
+	put_user(d_ino, &dirent->d_ino);
 	put_user(offset, &dirent->d_offset);
 	put_user(namlen, &dirent->d_namlen);
 	copy_to_user(dirent->d_name, name, namlen);

arch/powerpc/kernel/sys_ppc32.c

@@ -69,16 +69,20 @@ struct readdir_callback32 {
 };
 
 static int fillonedir(void * __buf, const char * name, int namlen,
-		                  off_t offset, ino_t ino, unsigned int d_type)
+		                  off_t offset, u64 ino, unsigned int d_type)
 {
 	struct readdir_callback32 * buf = (struct readdir_callback32 *) __buf;
 	struct old_linux_dirent32 __user * dirent;
+	ino_t d_ino;
 
 	if (buf->count)
 		return -EINVAL;
+	d_ino = ino;
+	if (sizeof(d_ino) < sizeof(ino) && d_ino != ino)
+		return -EOVERFLOW;
 	buf->count++;
 	dirent = buf->dirent;
-	put_user(ino, &dirent->d_ino);
+	put_user(d_ino, &dirent->d_ino);
 	put_user(offset, &dirent->d_offset);
 	put_user(namlen, &dirent->d_namlen);
 	copy_to_user(dirent->d_name, name, namlen);
@@ -120,15 +124,20 @@ asmlinkage long ppc32_select(u32 n, compat_ulong_t __user *inp,
 
 int cp_compat_stat(struct kstat *stat, struct compat_stat __user *statbuf)
 {
+	compat_ino_t ino;
 	long err;
 
 	if (stat->size > MAX_NON_LFS || !new_valid_dev(stat->dev) ||
 	    !new_valid_dev(stat->rdev))
 		return -EOVERFLOW;
 
+	ino = stat->ino;
+	if (sizeof(ino) < sizeof(stat->ino) && ino != stat->ino)
+		return -EOVERFLOW;
+
 	err  = access_ok(VERIFY_WRITE, statbuf, sizeof(*statbuf)) ? 0 : -EFAULT;
 	err |= __put_user(new_encode_dev(stat->dev), &statbuf->st_dev);
-	err |= __put_user(stat->ino, &statbuf->st_ino);
+	err |= __put_user(ino, &statbuf->st_ino);
 	err |= __put_user(stat->mode, &statbuf->st_mode);
 	err |= __put_user(stat->nlink, &statbuf->st_nlink);
 	err |= __put_user(stat->uid, &statbuf->st_uid);

arch/s390/kernel/compat_linux.c

@@ -357,11 +357,16 @@ asmlinkage long sys32_ftruncate64(unsigned int fd, unsigned long high, unsigned
 
 int cp_compat_stat(struct kstat *stat, struct compat_stat __user *statbuf)
 {
+	compat_ino_t ino;
 	int err;
 
 	if (!old_valid_dev(stat->dev) || !old_valid_dev(stat->rdev))
 		return -EOVERFLOW;
 
+	ino = stat->ino;
+	if (sizeof(ino) < sizeof(stat->ino) && ino != stat->ino)
+		return -EOVERFLOW;
+
 	err = put_user(old_encode_dev(stat->dev), &statbuf->st_dev);
 	err |= put_user(stat->ino, &statbuf->st_ino);
 	err |= put_user(stat->mode, &statbuf->st_mode);

arch/sparc/kernel/sys_sunos.c

@@ -325,21 +325,25 @@ struct sunos_dirent_callback {
 #define ROUND_UP(x) (((x)+sizeof(long)-1) & ~(sizeof(long)-1))
 
 static int sunos_filldir(void * __buf, const char * name, int namlen,
-			 loff_t offset, ino_t ino, unsigned int d_type)
+			 loff_t offset, u64 ino, unsigned int d_type)
 {
 	struct sunos_dirent __user *dirent;
 	struct sunos_dirent_callback * buf = __buf;
+	unsigned long d_ino;
 	int reclen = ROUND_UP(NAME_OFFSET(dirent) + namlen + 1);
 
 	buf->error = -EINVAL;	/* only used if we fail.. */
 	if (reclen > buf->count)
 		return -EINVAL;
+	d_ino = ino;
+	if (sizeof(d_ino) < sizeof(ino) && d_ino != ino)
+		return -EOVERFLOW;
 	dirent = buf->previous;
 	if (dirent)
 		put_user(offset, &dirent->d_off);
 	dirent = buf->curr;
 	buf->previous = dirent;
-	put_user(ino, &dirent->d_ino);
+	put_user(d_ino, &dirent->d_ino);
 	put_user(namlen, &dirent->d_namlen);
 	put_user(reclen, &dirent->d_reclen);
 	copy_to_user(dirent->d_name, name, namlen);
@@ -406,19 +410,23 @@ struct sunos_direntry_callback {
 };
 
 static int sunos_filldirentry(void * __buf, const char * name, int namlen,
-			      loff_t offset, ino_t ino, unsigned int d_type)
+			      loff_t offset, u64 ino, unsigned int d_type)
 {
 	struct sunos_direntry __user *dirent;
 	struct sunos_direntry_callback *buf = __buf;
+	unsigned long d_ino;
 	int reclen = ROUND_UP(NAME_OFFSET(dirent) + namlen + 1);
 
 	buf->error = -EINVAL;	/* only used if we fail.. */
 	if (reclen > buf->count)
 		return -EINVAL;
+	d_ino = ino;
+	if (sizeof(d_ino) < sizeof(ino) && d_ino != ino)
+		return -EOVERFLOW;
 	dirent = buf->previous;
 	dirent = buf->curr;
 	buf->previous = dirent;
-	put_user(ino, &dirent->d_ino);
+	put_user(d_ino, &dirent->d_ino);
 	put_user(namlen, &dirent->d_namlen);
 	put_user(reclen, &dirent->d_reclen);
 	copy_to_user(dirent->d_name, name, namlen);

arch/sparc64/kernel/sys_sparc32.c

@@ -337,12 +337,17 @@ asmlinkage long sys32_ftruncate64(unsigned int fd, unsigned long high, unsigned
 
 int cp_compat_stat(struct kstat *stat, struct compat_stat __user *statbuf)
 {
+	compat_ino_t ino;
 	int err;
 
 	if (stat->size > MAX_NON_LFS || !old_valid_dev(stat->dev) ||
 	    !old_valid_dev(stat->rdev))
 		return -EOVERFLOW;
 
+	ino = stat->ino;
+	if (sizeof(ino) < sizeof(stat->ino) && ino != stat->ino)
+		return -EOVERFLOW;
+
 	err  = put_user(old_encode_dev(stat->dev), &statbuf->st_dev);
 	err |= put_user(stat->ino, &statbuf->st_ino);
 	err |= put_user(stat->mode, &statbuf->st_mode);